It does look a lot better in that with with RT on, strikingly so. Very subtle but powerful.
More ‘life like’.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
The thread which sometimes talks about RDNA2
- Thread starter CAT-THE-FIFTH
- Start date
- Status
More ‘life like’.
Because the examples from DF are not cherry picked they blindly stopped the game from time to time and took some footage, it's not like they searched the best examples of RT usage in the game. 
I'll take native resolution without RT vs DLSS+RT any time. The game looks good enough without RT, too bad it is a big gameplay mess atm.
I'll take native resolution without RT vs DLSS+RT any time. The game looks good enough without RT, too bad it is a big gameplay mess atm.
Where RT really does look better in CP2077 is on glass surfaces. So you do get real time reflections of the world around you in glass etc. Though to be honest it then comes down to what is more important.
CP on high/medium at 4K on a 2080 with DLSS on balanced = Avg FPS of around 55
Same settings but with RT on medium and the FPS tanked down to 26 FPS on the same scene.
So DLSS 2.0 makes a 2080 a viable 4K card with VRR, but it doesn't make it a 4K with RT card. At least not in CP2077 unless you are prepared enable cateract mode (AKA DLSS on anything with the word performance in it). So if you are playing on 1440p then the 2080 becomes a decent RT enabled GPU with the right settings.
CP on high/medium at 4K on a 2080 with DLSS on balanced = Avg FPS of around 55
Same settings but with RT on medium and the FPS tanked down to 26 FPS on the same scene.
So DLSS 2.0 makes a 2080 a viable 4K card with VRR, but it doesn't make it a 4K with RT card. At least not in CP2077 unless you are prepared enable cateract mode (AKA DLSS on anything with the word performance in it). So if you are playing on 1440p then the 2080 becomes a decent RT enabled GPU with the right settings.
But as observers the large majority of people don't actually know what the accurate and right reflections and light paths are - only some professionals maybe know.
So still, there is no need for this error correction because it's all relative in the observers' eyes. Classic rendering can achieve the same end picture at much cheaper transistors cost - it's all about adjustments.
Well my 90+ hours of gaming on CP2077 has been fun and I personally think it looks fantastic with RT and hopefully AMD get their version out soon, so AMD users can run it. No bias from me and genuinely want all PC gamers to have a great experience in games and for me, RT does add another level of realism. I remember many years ago watching cut scenes in games thinking "I wish that was how the game looked" (using FFVII as an example) and now we have that, I can't wonder how long before games look like movies and with graphical advancements, not that long I hope.
Anyways, I hope you all have a good Christmas and get everything you want. We all love our hardware and I look forward to lots of petty bickering in 2021
Anyways, I hope you all have a good Christmas and get everything you want. We all love our hardware and I look forward to lots of petty bickering in 2021
Yup, realistic lighting, shadows, transparencies and reflections rendered in real-time are some of the key ingredients to next-level visual realism in games. That's just a statement of fact.
Anyways, I hope you all have a good Christmas and get everything you want. We all love our hardware and I look forward to lots of petty bickering in 2021
Merry Christmas matey !
Because the examples from DF are not cherry picked they blindly stopped the game from time to time and took some footage, it's not like they searched the best examples of RT usage in the game.
I'll take native resolution without RT vs DLSS+RT any time. The game looks good enough without RT, too bad it is a big gameplay mess atm.
Point is cherrypicking examples that dont show big differences is disingenous.
‘Cherrypicking’ examples that show big differences isn’t, because unlike the previous scenario no one is acting like there are big differences in every scenario but the others are acting like the ‘no big diff’ is the norm.
See the problem?!
That’s where their whole argument ( as if they ever had one ) falls apart.
Good
The only thing that looks better is the water - it looks more like it should be.
In all other aspects - I don't like the graphics of this game - for me Crysis 3 looks better and it's years earlier.
The human models look like doll, the physics level is extraordinarily low, everything looks like a cheap animation from 2005.
Putting ray-tracing effects on this and playing with the lights just for the sake of it (or for the sale of it) is like putting lipstick on a pig. Quite literally.
Extremely poor execution.
Might look good in the kids' eyes but this is not serious.
Just wow
I think we are beyond still images for comparison now anyway. Yeah use them if you want to dive into the detail differences after but what matters is how it looks in motion. People also say its a slow paced game which it is on the whole but action sequences can get a bit hectic and based on your characters build can also be fast moving too. Lets be honest in the middle of a fire fight you aren't going to be worried about whether a shadow falling on a nut on an air conditioning unit 30 metres away is correct or not...
He sure did spend a lot of time talking about RT shadows, according to certain fanboys, RT shadows is useless.
On a more serious note, ECH pictures aren't fake (AFAIK), so clearly some scenes benefit more/are more noticeable than others. But that may be down to time spent working on the raster scene and baking all the lights in.
The real argument is whether is is worth performance hit. In that 25 minute video there were 3 scenes that I feel RT made a significant enough difference to where i may consider turning it on. The water scene, the car scene, and the under the bridge scene. However the under the bridge scene, seems more like it was constructed to prove a point rather than something that would happen in normal gameplay. The car scene could be achieved with raster but it is the time argument.
People know this which is why they consider buying Ampere for its RT ability to be pretty pointless. If it is struggling to run CP2077 then it stands no chance in future, unless you buy a 3090 and run a 1080p monitor.
I'll say it again, people will jump on board when the performance hit to visual quality increase is worth it. This will most likely be when you can get 3090 RT performance for £300. Till then it will simply be a "oh look, that's nice" then turn it off feature.
Soldato
- Joined
- 14 Aug 2009
- Posts
- 3,251
Because the examples from DF are not cherry picked they blindly stopped the game from time to time and took some footage, it's not like they searched the best examples of RT usage in the game.
I'll take native resolution without RT vs DLSS+RT any time. The game looks good enough without RT, too bad it is a big gameplay mess atm.
You're missing the point. When the scenes look similar, then you don't lose anything in terms of visual quality while using RT, but when you move into places where RT stands out, you gain on top of the simple rasterization process.
In terms of performance is costly, indeed, but that's true with a lot of new tech. Plus, even with rasterization, on a RTX2080 it cannot hold 60fps locked in 1080p. Rasterization is getting expensive as well.
Using DLSS on Optimal is just fine; like I've said previously, the problem is with using other aggressive settings like ultra performance - which I would probably still us IF the whole game would have been path traced. DLSS aggressive settings can introduce shimmering in certain scenarios (fences, reflective surfaces like roads when is night and it rains and probably others).
True and if you don't have a day / night cycle, just a locked time of day, you can bake it in. Plus it also helps that playing for so long in the "fake" environment, we grew accustomed to it. Still, having a proper reflection without artifacts and proper lighting scattering and stuff can improve thing further.
Just the simple stuff of actually having shadows from every light in the scene would be a BIG step forward.
What benchmarks are you looking at?
Edit: Also it is a good thing that, that performance is now around £300.
Soldato
- Joined
- 14 Aug 2009
- Posts
- 3,251
You don't lose anything rendering at 720p and upscaling? OK, then. Enjoy your blurry mess with RTX ON
What about examples that show big NEGATIVE differences with RT enabled? I guess you'd like to swiftly brush those under the carpet...Point is cherrypicking examples that dont show big differences is disingenous.
‘Cherrypicking’ examples that show big differences isn’t, because unlike the previous scenario no one is acting like there are big differences in every scenario but the others are acting like the ‘no big diff’ is the norm.
See the problem?!
That’s where their whole argument ( as if they ever had one ) falls apart.
Good
The RTX shadows on the right are hopeless. Complete loss of detail.
But CP runs in digits on Nvidia cards too and yet that doesn't stop you to praise them and their great RT performance. Really open CP and try to run the game in 4k native with RT and see if you can get 10 FPS.
Why would AMD be afraid with that performance? It would be almost as good as Nvidia, close to 0.
Yet I have been running CP2077 since launch at 1440p with RT Phsyco, DLSS Quality and everthing else maxed with a 3080 and an ancient 3770k with 32GB DDR3 and SATA3 SSDs and getting 45 - capped 60 FPS. If I was using a 4k panel, I'd be using a 3090 and DLSS performance. bringing it back to the 60FPS mark.
The RTX shadows are more accurate even in your worst-case scenario example. They make no sense in the non-RT shot if you actually take a minute to think about them. For example, what on earth is supposed to be causing those pipes to cast those shadows on the wall behind them in the non-RT shot? The sun clearly isn't shining from that direction. I notice that that's also the 'Ultra' RT preset without GI enabled too, which makes a large difference to outdoor shadowing. What does it look like on Psycho?
It's funny how to some showing screens where RT doesn't make a difference is "cherry picking" but they're all too happy to pretend their 1080p upscaled image is really 4K, even though DLSS can't touch the RT reflections and keeps them low res.
How many of these RT "enthusiasts" have a 3090 and are playing at 20 fps because RT just looks so good? And even if we delude ourselves into thinking DLSS is special & gets near-native (it doesn't), how special is it to barely go above 30 fps?! And that's WITH a tuned 5950x which is necessary because of how CPU demanding RT also is. lmao
How many of these RT "enthusiasts" have a 3090 and are playing at 20 fps because RT just looks so good? And even if we delude ourselves into thinking DLSS is special & gets near-native (it doesn't), how special is it to barely go above 30 fps?! And that's WITH a tuned 5950x which is necessary because of how CPU demanding RT also is. lmao
at 1080p?I'm not looking at any bechmarks, I'm playing it!
- Status